A Question and Answer System for Program Comprehension based on Encoder-Decoder Recurrent Neural Networks

Computer programming is an innovative cognitive tool that has transformed modern society. Two of the most widely used programming languages for web development are java and python. Therefore, source code comprehension of those programming is considered as an essential part and time-consuming task during software maintenance process. To support code comprehension process, the question and answer systems based on program comprehension have proposed. Recurrent Neural Network (RNN) based sequence-to-sequence model is one of the most commonly researched models to implement artificial intelligence question and answer system. However, it is not being applied widely in question and answer system for code comprehension. This system will learn using neural network where it uses bidirectional RNN as encoder and Luong Attention RNN as decoder. This system proposed a question and answer system to provide source code comprehension with encoder-decoder RNNs based on the code comprehension dataset: CodeQA.